Click on the links below to access all the ArticlePlus for this article.
Please note that ArticlePlus files may launch a viewer application outside of your web browser.
In 2001, the United Nations General Assembly Special Session on HIV/AIDS (UNGASS) passed a declaration of commitment to reduce global mother-to-child transmission (MTCT) of HIV by 20% by 2005 and by 50% by 2010.1 The commitment stressed that the goal should be achieved by implementing a range of public health interventions guided by nationally coordinated strategies. These include ensuring that 80% of pregnant women with access to antenatal care are provided with HIV prevention services, including voluntary counseling and testing (VCT) and antiretroviral therapy (ART).
Prolonged efficacy of ART for prevention of mother-to-child transmission (PMTCT) of HIV has been confirmed in clinical studies in non-breast-feeding populations in middle- and high-income countries.2-5 Efficacy is typically somewhat lower in breast-feeding populations in Africa.6-12 In developing countries, the effectiveness of these ART regimens at the population level remains unknown,13 because empiric studies to evaluate their effectiveness are costly. In addition, where attempts have been made to measure the effectiveness of these interventions in real-life settings, sample sizes have been too small for the findings to be generalized.14,15 Mathematic models, parameterized using local data, can be a useful tool to help explore the likely effectiveness of these interventions and to provide information to improve program impact.16,17
With HIV prevalence in pregnant women attending antenatal clinics (ANCs) recorded at 32.1% in 2000 and at 23.9% in 2004,18 the risk of MTCT of HIV in Zimbabwe, as in many other countries in sub-Saharan Africa, is extremely high. In recognition of this and in accordance with the UNGASS recommendations, the Zimbabwe Ministry of Health and Child Welfare designed and implemented a PMTCT program in 2001 based on provision of single-dose nevirapine (NVP) to pregnant women and their infants.19 Between 2002 and 2005, the program was scaled up, with increasing numbers of health facilities being registered and equipped to provide PMTCT services.
This article aims to assess (1) the extent of MTCT of HIV in Zimbabwe from 1980 to 2005, (2) the impact of the current single-dose NVP PMTCT program in Zimbabwe in 2005 (the most recent year for which complete data are available), and (3) how the impact of PMTCT could be maximized in similar high-seroprevalence settings.
Decision Analysis Model
A decision analysis model16,17 representing the experience of successive cohorts of pregnant women enrolling at ANCs at 5 months of pregnancy was developed and analyzed using spreadsheet-based Precision Tree 1.0 and @Risk 4.5 software (Palisade Corporation, Ithaca, NY).20
In the model, a proportion of each cohort of pregnant women (i) is assumed to be infected with HIV at the first ANC visit. This proportion is assumed to be the same (i) in women who do and do not have access to VCT for PMTCT.21 Women who are not infected with HIV when enrolled for ANC care can seroconvert during the remaining term of the pregnancy (at a rate s1) or during breast-feeding (s2), risks that are assumed to be independent of VCT uptake. The proportion of women (β) choosing to breast-feed their infants is assumed to be independent of HIV and VCT status. Vertical infection can occur in utero, intrapartum, or postpartum (by means of breast-feeding). The risk of vertical transmission can be set to be higher for women who seroconvert during pregnancy or breast-feeding.
The coverage was defined as the proportion (a) of each cohort of women with access to VCT for PMTCT. A proportion (va) of the women with access to VCT for PMTCT accept VCT independent of their HIV status. Women found to be infected with HIV when presenting for VCT are offered ART to reduce MTCT (in the case of Zimbabwe, a standard single-dose NVP regimen), and a fraction (ta) of these accept.
In the model, key parameters such as HIV prevalence in pregnant women and coverage and acceptance of PMTCT services can be varied over time and/or by geographic location by resetting the parameters for successive cohorts of pregnant women.
Model parameter estimates and uncertainty ranges were based on data from published literature, ongoing studies, and service statistics from Zimbabwe's national intervention programs.
Estimates of HIV prevalence in pregnant women (i) (Fig. S1; Supplemental materials are available via the Article Plus feature at www.jaids.com. You may locate this article, then click on the Article Plus link on the right) were derived from the Epidemiology Projection Program (EPP)-based national estimates for 1980 to 2005,22 which, in turn, were based on the national ANC surveillance data (See Table S1).23,24
Most women in Zimbabwe breast-feed their infants for 15 to 18 months.25,26 In the model, we assume that 95% (range: 90% to 100%) of women who give birth breast-fed their infants for 24 months at most and that the remainder did not breast-feed. The latter allows for the professional urban women and women who are ill, who do not breast-feed at all or do so only for short periods.
HIV incidence over 12 months in childbearing women was assumed to range between 1.8% and 3%, based on a previous EPP fit27 and to be constant during pregnancy and breast-feeding. The risk of maternal seroconversion is assumed to occur for 4 months during pregnancy (from the fifth month of gestation until delivery) and for 6 months during breast-feeding. To calculate MTCT risk in recently seroconverted women (s1 and s2), we assumed that 15%, 50%, and 35% of all vertical infections occur in utero (conception to just before labor), intrapartum (during labor), and postpartum (during or through breast-feeding), respectively.28 Women who seroconverted between the first HIV test (or fifth month of pregnancy) and delivery were assumed to do so at 7 months of pregnancy, thereby exposing their infants to HIV infection for 2 months in utero and intrapartum. Women infected in the year after delivery were assumed to seroconvert at 6 months; thus, infants of seroconverting breast-feeding women were assumed to be exposed to HIV infection for 18 months.
The overall risk of MTCT in untreated and breast-fed infants (t1β) is estimated to be 0.32 (range: 0.20 to 0.40; Table 1).29,30 In untreated and non-breast-fed infants, the cumulative transmission probability is estimated to be 0.21 (range: 0.13 to 0.30). The transmission probability in recently seroconverted women who breast-feed (t2β) is estimated to be 0.43 (range: 0.38 to 0.48), 35% (range: 20% to 50%) more than in women infected in the past,31-33 whereas in women who seroconvert during pregnancy but do not breast-feed (t2nβ), the probability of MTCT is 0.23 (range: 0.17 to 0.30). For those who breast-feed and seroconvert during breast-feeding, the transmission probability is taken to be 0.17 (range: 0.13 to 0.25).
The efficacy of a given PMTCT intervention from time of administration to the mother and infant up to the age of 24 months is reflected in the model by adjusting the transmission probability by a factor of ξ, which varies by form of treatment (see Table 1). The efficacy of single-dose NVP (ξnvp) is taken to be 0.41 (95% confidence interval [CI]: 0.19 to 0.59),8 whereas the efficacy of long-course maternal zidovudine (ZDV) combined with NVP and infant NVP is estimated to be 0.90 (95% CI: 0.70 to 0.97) (Dube S, Boily MC, Hader S, et al. The effect of antiretroviral therapy on mother-to-child transmission of HIV-1: a meta-analysis. 2006. Unpublished.).
The proportion of pregnant women with access to PMTCT services (coverage) in a given year (a) was taken as the total number of women who attended ANCs offering PMTCT divided by the total number of women who gave birth. Demographic and Health Survey and United Nations population reports indicate that approximately 400,000 births occurred each year in Zimbabwe since the late 1990s, although from 1980 up to the mid-1990s, the number of births varied between 360,000 and 384,000.35,36 In 2002, when the PMTCT program began in earnest, approximately 46,000 women sought antenatal care at clinics offering PMTCT services. This figure increased to 98,000 in 2003, 116,000 in 2004, and 196,000 in 2005.19,37,38 Therefore, for example, PMTCT coverage increased from 10% (range: 5% to 15%) in 2002 to 50% (range: 40% to 60%) in 2005. PMTCT acceptance rates (va) assumed in the model were also based on service statistics.19,37,38 In 2005, 65% (range: 60% to 70%) of pretest-counseled women accepted HIV testing, and of those who tested positive, the same proportion took up the single-dose NVP for PMTCT. Program coverage is higher in urban areas than in nonurban areas (Fig. S2; Supplemental materials are available via the Article Plus feature at www.jaids.com. You may locate this article, then click on the Article Plus link on the right). Data on trends in acceptance rates over time disaggregated by location were not available; thus, the overall trend was assumed to have occurred in all residential areas. Similar patterns of coverage and acceptance of PMTCT programs have been seen in other African countries,13,39-41 with a review covering 12 countries reporting VCT acceptance levels of between 40% and 80%.19,37,39
The model analysis first examined a “no-intervention scenario” that estimated (1) the proportion and number of infants infected each year from 1980 to 2005, (2) the cumulative number of infants infected from the beginning of the epidemic to 2005, and (3) the proportions of infections attributable to breast-feeding and seroconversion in 2005 in the absence of PMTCT. Then, the proportion and numbers of infections averted by the PMTCT program between 2001 and 2005 were estimated by comparing the intervention scenario with the no-intervention scenario.
Estimates were also made by place of residence-rural, urban and “other” areas (growth points, national parks, army camps, and commercial farms or estates)-assuming that 58%, 32%, and 10% of the total population resides in urban, rural and other areas, respectively,42 and using area-specific HIV prevalence and intervention coverage estimates (see Figs. S1, S2; Supplemental materials are available via the Article Plus feature at www.jaids.com. You may locate this article, then click on the Article Plus link on the right).
The potential additional population-level effects of alternative regimens to the single-dose NVP were explored by setting the regimen efficacy parameter to the alternative regimen efficacies measured in trials and repeating the simulation using 2005 as the referent year. The impact of the current NVP regimen was compared with that of long- and short-course ZDV monotherapy,2,3,12,43 combined long-course ZDV+NVP,4 and a short-course ZDV + lamivudine (3TC) regimen (arm A).11 In the final analysis, however, less efficacious regimens were dropped and the analysis was focused on comparing the impact of the highly efficacious combined and long-course regimen of ZDV+NVP and the currently used single-dose NVP regimen. The impact of using these alternative regimens was then estimated under increasing levels of program coverage and acceptance.
To reflect the uncertainty in parameter assumptions on model predictions, Latin Hypercube sampling techniques44,45 were used over predefined parameter ranges with a uniform probability distribution. The parameter ranges were chosen to represent the lowest and highest possible values from available data. When direct data were unavailable, best estimates were extrapolated using simple calculations (shown in Table 1). To identify important programmatic and epidemiologic parameters and to explore the sensitivity of key model predictions to uncertainty in parameter estimates, univariate and multivariate uncertainty and sensitivity analyses were carried out, again using 2005 as a referent year. In the uncertainty analysis, parameters were varied within the predefined ranges (shown in Table 1), whereas for the sensitivity analysis, they were varied over a wider range that reflects extreme minimum and maximum values for the different parameters. In the multivariate uncertainty and sensitivity analyses, the independent impact of each specified model parameter on model output was assessed by examining the adjusted regression coefficients.20
Finally, 2-way sensitivity analysis of program coverage and regimen efficacy were carried out to investigate the levels of coverage and regimen efficacy needed to achieve 20% and 50% reductions in MTCT as espoused by the UNGASS declaration.
Pediatric HIV Epidemiology and the Impact of PMTCT in Zimbabwe
We estimate that pediatric HIV infections attributable to MTCT were <1% before 1987 and rose exponentially from then on as more adults became infected. Following trends in HIV infection in adults (see Figs. S1, S3; Supplemental materials are available via the Article Plus feature at www.jaids.com. You may locate this article, then click on the Article Plus link on the right), infections in infants peaked at 9.8% (range: 7.4% to 12.1%) in the mid-1990s (Fig. 1A) and declined from 8.2% (range: 6.0% to 10.7%) in 2000 to 6.9% (range: 5.3% to 8.7%) in 2003, 6.5% (range: 5.0% to 8.1%) in 2004, and 6.2% (range: 4.9% to 8.9%) in 2005. From the beginning of the HIV epidemic up to 2005, we estimate that of approximately 10 million children born, 504,000 (range: 362,000 to 666,000) infants were vertically infected with HIV (the period between 2002 and 2005 includes PMTCT intervention). Annual infections peaked at 39,000 (range: 30,000 to 49,000) in 1995 before decreasing to 32,000 (range: 25,000 to 40,000) in 2000 and to 24,500 (range: 19,000 to 32,000) in 2005 (see Fig. 1B).
Vertical transmission peaked at different times in the 3 geographic areas. Infant infections in urban areas peaked first at approximately 12% in the mid-1990s and show a greater reduction than in rural areas as a result of PMTCT (see Fig. 1C). Subsequent peaks occurred in rural areas (8%) and then in other areas (12%) in the late 1990s. In 2005, we estimated that 7.2% (range: 5.1% to 9.3%), 5.1% (range: 3.7% to 6.5%), and 8.1% (range: 6.0% to 10.4%) of infants born in urban, rural, and other areas, respectively, were vertically infected with HIV. Infants born in the other areas were estimated to be at higher risk of infection compared with those born in rural or urban areas because of higher HIV prevalence in these areas. The much larger size of the rural population means that a greater number of pediatric infections occur in rural areas than in urban or other areas, however. Thus, we estimate that in 2005, approximately 11,500 (range: 8600 to 15,000) infants were vertically infected with HIV in rural areas compared with 8700 (range: 6300 to 11,400) in urban areas and 3400 (range: 2300 to 4000) in other areas. Figure 2 indicates that without PMTCT, the proportion of infants infected with HIV in Zimbabwe would have declined as a result of the decrease in adult HIV prevalence, from 8.2% (range: 6.0% to 10.7%) in 2000 to 6.8% (range: 4.2% to 10.9%) in 2005. In 2005, 32% (range: 26% to 44%) and 4.0% (range: 2.7% to 6.2%) of infections were attributable to breast-feeding and maternal seroconversion, respectively, and the PMTCT program reduced infant infections by 8.8% (range: 5.5% to 12.1%).
By comparing intervention and no-intervention model scenarios (Fig. 2A), we estimate that the PMTCT program reduced infant infections by 1.5% (range: 1.0% to 2.8%), 2.8% (range: 2.2% to 6.5%), 4.4% (range: 2.8% to 5.6%), and 8.8% (range: 5.5% to 12.1%) in 2002, 2003, 2004, and 2005, respectively. By the end of 2005, a cumulative 4600 (range: 3900 to 7800) infections were averted as a direct result of the PMTCT program. Given the same intervention coverage and acceptance levels, the number of infections averted could have been doubled had the more efficacious NVP+ZDV4 regimen been implemented (see Fig. 2A). With current coverage and acceptance rates, the proportion of infants infected would be reduced by more than a third if infections during breast-feeding could be avoided; for example, through breast-feeding avoidance among infected mothers.
Assuming linear increases over time in program coverage (from 50% to 100%) and intervention acceptance (from 65% to 100%) and that HIV prevalence in pregnant women declines in line with national estimates, by 2010, the current single-dose NVP program can be estimated to have prevented a cumulative 16,300 infections compared with 35,200 infections if an ZDV+NVP regimen had been used from the start. There is negligible difference in impact if alternative regimens of short-course ZDV (Abidjan and DITRAME)6,9 or short-course ZDV+3TC (Petra study)11 were used in place of the NVP regimen, however. This is because the efficacies of these regimens are too low to have a substantial impact at the population level.
The fraction of vertical transmissions averted by the current program is limited by a combination of low VCT and antiretroviral (ARV) coverage and acceptance (see Figs. 2B, C). Given the current low levels of program acceptance, expanding the program to maximum coverage would have little impact unless acceptance is also increased simultaneously (see Fig. 2C), even if a more efficacious regimen such as ZDV+NVP were used. Indeed, program impact is less sensitive to increases in regimen efficacy when coverage and acceptance levels are low. Regimen efficacy becomes increasingly important as coverage and acceptance of VCT and ART rise concurrently (but not when only coverage or acceptance increases).
Uncertainty and Sensitivity Analysis
The results of the univariate and multivariate sensitivity analysis for 2005 are shown in Figure 1. In this analysis, a regression value of 0 indicates that there is no significant relation between the input and the output, whereas a regression value of 1 or −1 indicates a +1 or −1 SD change in the output for a 1-SD change in the input.20 The uncertainty analysis results suggest that given the predefined ranges of VCT and ART acceptance initially explored, the predicted proportion of MTCT in 2005 is most sensitive to uncertainty in HIV prevalence, baseline MTCT risk, regimen efficacy, and breast-feeding. Program coverage, VCT acceptance, and HIV incidence have relatively less influence (see Fig. 3A). In the broader sensitivity analysis, when the parameters are varied over a wider range, breast-feeding becomes more important, followed by background HIV prevalence and MTCT risk (see Fig. 3B).
Despite progress in the rollout of PMTCT programs, it seems that Zimbabwe was not able to achieve the UNGASS “goal” to reduce MTCT by 20% by the end of 2005 based on the performance of the PMTCT program. Based on current prevalence levels, to achieve a 20% reduction in MTCT using the present NVP regimen, program coverage and acceptance would have had to have been increased to at least 80% and 95%, respectively (Table S2; Supplemental materials are available via the Article Plus feature at www.jaids.com. You may locate this article, then click on the Article Plus link on the right). Furthermore, to reduce MTCT by 50% (a 2010 UNGASS target), the national program in Zimbabwe must increase PMTCT coverage to 90%, increase acceptance of VCT and ART to 100%, and use a regimen with at least 60% efficacy (Table S3; Supplemental materials are available via the Article Plus feature at www.jaids.com. You may locate this article, then click on the Article Plus link on the right).
This article highlights the scale of MTCT of HIV and, despite progress in the rollout of ART, the limited impact of PMTCT interventions in Zimbabwe. From the beginning of the epidemic up to 2005, a cumulative 504,000 (range: 362,000 to 666,000) infants could have been infected with HIV through MTCT; in 2005 alone, approximately 24,500 (range: 19,000 to 32,000) infants may have been infected this way. The official Zimbabwe National HIV/AIDS Estimate for 2005 of 26,600 new infections in the age group from 0 to 14 years22 falls within our range of estimates.
The sharp declines in MTCT estimated to have occurred between 2000 and 2005 are mainly attributable to the concurrent decline in HIV prevalence in adults and less so to the rollout of PMTCT services. Our results suggest that the PMTCT program introduced in 2001 has had a modest effect on MTCT. In 2005, low PMTCT coverage and acceptance resulted in the program averting approximately 8.8% (range: 5.5% to 12.1%) of all potential MTCT infections in infants.
Despite the optimism brought about by the new interventions and progress in the right direction, vertical transmission of HIV remains a major public health problem for Zimbabwe and other similar resource-poor countries with high HIV prevalence. Thus, efforts should be intensified, especially if UNGASS targets are to be met. The current PMTCT program has had only a modest impact on MTCT of HIV in Zimbabwe, well below UNGASS goals, principally because of low coverage and acceptance of services and low efficacy of the widely used single-dose NVP regimen. As such, the program priority should be to increase coverage and VCT and ART acceptance levels to >75%. Substituting the current single-dose NVP regimen with a more efficacious regimen becomes more beneficial once higher program coverage levels are reached. These results assume that the current declines in adult HIV prevalence are unrelated to UNGASS-related intervention activities, however. The effect of the latter could be seen to be higher if we assume that the declines in adult HIV prevalence can be attributed to UNGASS-related behavior change interventions, as is suggested by a recent epidemiologic review.46
The impact of PMTCT services in Zimbabwe is likely to increase in the future. Efforts continue to be made to scale up the coverage of PMTCT services. VCT services generally are also being scaled up, in part, to facilitate the highly active antiretroviral therapy (HAART) program, and a change in national policy has been made from client-initiated (“opt-in”) to provider-initiated (“opt-out”) VCT in all health service settings, which should substantially increase acceptance of VCT and PMTCT.47 There could also be an increase in selective avoidance of breast-feeding in infected women who know their HIV status, leading to steeper declines in MTCT than predicted here. Wider rollout of HAART could increase maternal survival and reduce HIV-associated subfertility,48 which, in turn, would slow the rate of decline in the number of pediatric HIV infections.
These results highlight challenges facing PMTCT interventions in affected countries. Expanding access and acceptance to ANC-based PMTCT services is the primary key to increasing program effectiveness. Where PMTCT services are available, VCT acceptance is often low, however.13,39-41 Even for those who accept VCT and test HIV-positive in ANC settings, only half accept ART. Scale-up efforts thus must be accompanied by coordinated community mobilization efforts to increase acceptance of VCT and ARV therapy.
Currently, in Zimbabwe, model estimates suggest that breast-feeding contributes a substantial amount of infections, approximately 36% (range: 30% to 45%), where PMTCT interventions are available. As PMTCT coverage increases, the fraction of all infections attributable to breast-feeding is expected to rise further, because current regimens are more effective in reducing intrauterine infections. Nevertheless, recommending complete breast-feeding avoidance in resource-poor settings needs further investigation; it is complicated by competing risks of infant or child morbidity and mortality, because breast-feeding provides important protection against serious diseases that account for greater than two thirds of mortality in children younger than 5 years of age in these settings.49-51
At this stage of the epidemic, which is characterized by high HIV seroprevalence and decreasing incidence, seroconversion during pregnancy and breast-feeding contributes only marginally to infection in infants, with an estimated 4.0% (range: 2.7% to 6.2%; approximately 1200 cases in 2005) of infections occurring in recently seroconverted women. Nevertheless, programs should aim to prevent these infections by encouraging condom use in high-risk pregnant women, particularly because this may prevent infections in future pregnancies or in sexual partners.
Our estimates are based on a number of simplified model assumptions using the sometimes limited data available to estimate model parameters. HIV prevalence estimates were taken from the EPP-based national estimates. These estimates reflect HIV prevalence in all male and female adults in Zimbabwe, which may be slightly lower than among pregnant women.52 Nevertheless, the national estimate for 2005 (20.1%, range: 17.0% to 23.5%) was somewhat higher than HIV prevalence in all adults in a large-scale nationally representative survey carried out in Zimbabwe in the same year (18.1%).53 Furthermore, the use of wider uncertainty ranges for the parameters provides for any residual discrepancy. Breast-feeding duration may be shorter than we have assumed and exclusive breast-feeding more widely practiced,53 resulting in lower MTCT than estimated here. In real-life settings, such as the one evaluated in this article, the efficacy of the singe-dose NVP regimen may be lower than in clinical trials.13,54 We also assume that acceptance of VCT is similar in HIV-infected and uninfected women. Women who suspect they are infected may be more likely to seek and accept testing,55 but any difference seems likely to have been small before the widespread availability of ART.56
This relatively simple model provides useful clues on the extent of MTCT and the impact of PMTCT in Zimbabwe. In addition, model estimates of the likely population-level impact of PMTCT programs based on expected levels of coverage and acceptance could be used in the design of empiric evaluation studies (eg, in sample size calculations). The findings reported here suggest that empiric studies of program impact would need to be unfeasibly large to capture the population-level effect of the present program, given its current level of coverage and acceptance. More complex dynamic models would be needed to estimate the long-term impact of interrelated interventions, however, including VCT, PMTCT, and provision of ART to pregnant and breast-feeding HIV-infected women. These models are needed to analyze the contribution of HIV/AIDS and the impact of interventions on early childhood mortality and, particularly, to examine the effect of the epidemiologic context on the impact of infant feeding alternatives such as replacement feeding, exclusive breast-feeding, and different durations of breast-feeding.
In Zimbabwe, at present, HIV interventions are concentrated in urban areas because of the greater availability of more advanced health infrastructure and disproportionate investment by AIDS service organizations. At the national level, however, the impact of targeting the “easy-to-reach” urban populations is limited, because most pediatric HIV infections occur in the “hard-to-reach” and generally underserved rural populations. To reduce MTCT substantially in Zimbabwe and similar sub-Saharan African countries, greater efforts have to be made to improve PMTCT services in rural areas.
The strong interaction between program coverage, acceptance, and regimen efficacy found in this model calls for closer reflection on how intervention programs are scaled up in such settings. The results suggest that substantial investment should be channeled first to build capacity to expand coverage of PMTCT services and to mobilize communities to increase uptake of these services.57 This should then provide a platform for the introduction of more efficacious ART regimens that have a substantial population-level impact. Introducing highly efficacious but more complex regimens requires substantial infrastructural and human capacity development for health systems to cope with follow-up and monitoring activities. Therefore, further investment in the development of similarly simple but more efficacious regimens to improve on the single-dose NVP regimen could be considered. Furthermore, given its low efficacy and lack of benefit for maternal health13 and concerns about NVP resistance,58-60 more research needs to be conducted to establish the actual long-term gain of single-dose NVP as an intervention to reduce MTCT and its cost-effectiveness, so as to guide the prioritization of PMTCT interventions.
1. United Nations General Assembly Special Session. Action Guide for the United Nations Country Teams: Implementing the Declaration of Commitment on HIV/AIDS
. United Nations Special Session on HIV/AIDS. New York: United Nations General Assembly; 2001.
2. Connor EM, Sperling RS, Gelber R, et al. Reduction of maternal-infant transmission of human immunodeficiency virus type 1 with zidovudine treatment. Pediatric AIDS Clinical Trials Group Protocol 076 Study Group. N Engl J Med
3. Lallemant M. A trial of shortened zidovudine regimens to prevent mother-to-child transmission of human immunodeficiency virus type 1. Perinatal HIV Prevention Trial (Thailand) Investigators. N Engl J Med
4. Lallemant M, Jourdain G, Le Coeur S, et al. Single-dose perinatal nevirapine plus standard zidovudine to prevent mother-to-child transmission of HIV-1 in Thailand. N Engl J Med
5. Mandelbrot L, Landreau-Mascaro A, Rekacewicz C, et al. Lamivudine-zidovudine combination for prevention of maternal-infant transmission of HIV-1. JAMA
6. Dabis F, Msellati P, Meda N, et al. Six-month efficacy, tolerance, and acceptability of a short regimen of oral zidovudine to reduce vertical transmission
of HIV in breastfed children in Côte d'Ivoire and Burkina Faso; a double-blind placebo-controlled multicentre trial. Lancet
7. Guay LA, Musoke P, Fleming T, et al. Intrapartum and neonatal single-dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: HIVNET 012 randomised trial. Lancet
8. Jackson JB, Musoke P, Fleming T, et al. Intrapartum and neonatal single-dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: 18-month follow-up of the HIVNET 012 randomised trial. Lancet
9. Leroy V. Postnatal transmission of HIV-1 after a maternal short-course zidovudine peripartum regimen in West Africa: a pooled analysis of two randomised clinical trials. AIDS
10. Leroy V, Karon JM, Alioum A, et al. Twenty-four month efficacy of a maternal short-course zidovudine regimen to prevent mother-to-child transmission of HIV-1 in West Africa. AIDS
11. Petra Study Group. Efficacy of three short-course regimens of zidovudine and lamivudine in preventing early and late transmission of HIV-1 from mother to child in Tanzania, South Africa, and Uganda (Petra study): a randomised, double-blind, placebo-controlled trial. Lancet
12. Wiktor SZ, Ekpini E, Karon JM, et al. Short-course oral zidovudine for prevention of mother-to-child transmission of HIV-1 in Abidjan, Côte d'Ivoire: a randomised trial. Lancet
13. Quaghebeur A, Mutunga L, Mwanyumba F, et al. Low efficacy of nevirapine (HIVNET012) in preventing perinatal HIV-1 transmission in a real-life situation. AIDS
14. Colvin M, Chopra M, Doherty T, et al. Operational effectiveness of single-dose nevirapine in preventing mother-to-child transmission of HIV. Bull World Health Organ
15. Harvey K, Thame I. The Impact of a Programme to Prevent Mother-to-Child Transmission of HIV Disease: Transmission and Health-Seeking Behaviour Among HIV-Positive Mother-Child Pairs in Jamaica
. Bethesda, MD: Quality Assurance Project; 2004.
16. Weinstein MC, O'Brien B, Hornberger J, et al. Principles of good practice for decision analytic modeling in health-care evaluation: report of the ISPOR Task Force on Good Research Practices-Modeling Studies. Value Health
17. Weinstein MC, Toy EL, Sandberg EA, et al. Modeling for health care and other policy decisions: uses, roles, and validity. Value Health
18. Mahomva A, Greby S, Dube S, et al. HIV prevalence and trends from data in Zimbabwe, 1997-2004. Sex Transm Infect
. 2006;82(Suppl 1):i42-i47.
19. Ministry of Health and Child Welfare. Zimbabwe Programme for Prevention of Mother to Child Transmission of HIV (PMTCT): 2002 Annual Report
. Harare, Zimbabwe: Ministry of Health and Child Welfare; 2002.
20. Palisade Corporation. @Risk: Advanced Risk Analysis for Spreadsheets (A Guide)
21. Ministry of Health and Child Welfare. Needs Assessment for Prevention of Mother-to-Child Transmission of HIV in Zimbabwe
. Harare, Zimbabwe: Ministry of Health and Child Welfare; 1998.
22. Ministry of Health and Child Welfare. Zimbabwe National HIV and AIDS Estimates
. Harare, Zimbabwe: Ministry of Health and Child Welfare; 2006.
23. Ministry of Health and Child Welfare. National Survey of HIV and Syphilis Among Women Attending Antenatal Clinics in Zimbabwe
. Harare, Zimbabwe: Ministry of Health and Child Welfare; 2000.
24. Ministry of Health and Child Welfare. National Survey of HIV and Syphilis Among Women Attending Antenatal Clinics in Zimbabwe, 2003
. Harare, Zimbabwe: Ministry of Health and Child Welfare; 2004.
25. Iliff PJ, Piwoz EG, Tavengwa NV, et al. Early exclusive breastfeeding reduces the risk of postnatal HIV-1 transmission and increases HIV-free survival. AIDS
26. Gottlieb D, Shetty A, Mapfungautsi RM, et al. Infant feeding practices of HIV-infected and uninfected women in Zimbabwe. AIDS Patient Care STDS
27. Ministry of Health and Child Welfare. Zimbabwe National HIV and AIDS Estimates
. Harare, Zimbabwe: Ministry of Health and Child Welfare; 2003.
28. De Cock KM, Fowler MG, Mercier E, et al. Prevention of mother-to-child HIV transmission in resource-poor countries: translating research into policy and practice. JAMA
29. Iliff PJ, Piwoz EG, Tavengwa NV, et al. Early exclusive breastfeeding reduces the risk of postnatal HIV-1 transmission and increases HIV-free survival. AIDS
30. Nathoo KJRS, Zijenah LS, Kasule J, et al. Survival pattern among infants born to Human Immunodeficiency Virus type-1 infected mothers and uninfected mothers in Harare, Zimbabwe. Cent Afr J Med
31. John GC, Nduati RW, Mbori-Ngacha DA, et al. Correlates of mother-to-child human immunodeficiency virus type 1 (HIV-1) transmission: association with maternal plasma HIV-1 RNA load, genital HIV-1 DNA shedding, and breast infections. J Infect Dis
32. Semba RD, Kumwenda N, Hoover DR, et al. Human immunodeficiency virus load in breast milk, mastitis, and mother-to-child transmission of human immunodeficiency virus type 1. J Infect Dis
33. Jamieson DJ, Sibailly TS, Sadek R, et al. HIV-1 viral load and other risk factors for mother-to-child transmission of HIV-1 in a breast-feeding population in Côte d'Ivoire. J Acquir Immune Defic Syndr
35. Central Statistical Office. Zimbabwe Demographic and Health Survey
. Harare, Zimbabwe, 1999.
36. Population Division of the Department of Economic and Social Affairs of the United Nations Secretariat. World Population prospects: the 2004 revision. Available at: http://esa.un.org/unpp
. Accessed October 25, 2007.
37. Ministry of Health and Child Welfare. Zimbabwe Programme for Prevention of Mother to Child Transmission of HIV (PMTCT): 2004 Annual Report
. Harare, Zimbabwe: Ministry of Health and Child Welfare; 2004.
38. Ministry of Health and Child Welfare. Zimbabwe Programme for Prevention of Mother to Child Transmission of HIV (PMTCT): Service Data
. Harare, Zimbabwe: Ministry of Health and Child Welfare; 2006.
39. Buyse D, Nuwaha FN, Karlin T, et al. Prevention of mother to child transmission of HIV: from research to action. Presented at: XIVth International Conference on AIDS; Barcelona, 2002.
40. Temmerman M, Quaghebeur A, Mwanyumba F, et al. Mother-to-child HIV transmission in resource poor settings: how to improve coverage? AIDS
41. van't Hoog AH, Mbori-Ngacha DA, Marum LH, et al. Preventing mother-to-child transmission of HIV in Western Kenya: operational issues. J Acquir Immune Defic Syndr
42. Central Statistical Office. Census 2002: Zimbabwe National Report
. Harare, Zimbabwe, 2005.
43. Lallemant MJG, Le Coeur S, Kim S, et al. A trial of shortened zidovudine regimens to prevent mother-to-child transmission of human immuno-deficiency virus type 1. Perinatal HIV Prevention Trial (Thailand) Investigators. N Engl J Med
44. Stein M. Large sample properties of simulations using Latin hypercube sampling. Technometrics
. 1987;29:143-151. Available at: http://links.jstor.org/sici?sici=0040-1706(200002)42%3A1%3C55%3AACOTMF%3E2.0.CO%3B2-6
45. McKay MD, Beckman RJ, Conover WJ. A comparison of three methods for selecting values of output variables in the analysis of output from a computer code. Technometrics
. 1979;21:239-235. Available at: http://links.jstor.org/sici?sici=0040-1706
46. UNAIDS. Evidence for HIV Decline in Zimbabwe: A Comprehensive Review of the Epidemiological Data
. Geneva, Switzerland: UNAIDS; 2005.
47. Perez F, Zvandaziva C, Engelsmann B, et al. Acceptability of routine HIV testing (“opt-out”) in antenatal services in two rural districts of Zimbabwe. J Acquir Immune Defic Syndr
48. Zaba B, Gregson S. Measuring the impact of HIV on fertility in Africa. AIDS
. 1998;12(Suppl 1):S41-S50.
49. Jason JM, Nieburg P, Marks JS. Mortality and infectious diseases associated with infant feeding practices in developing countries. Pediatrics
50. Victoria CG. Infection and disease; the impact of early weaning. Food Nutr Bull
51. WHO. Effect of breastfeeding on infant and child mortality due to infectious diseases in less developed countries: a pooled analysis. WHO Collaborative Study Team on the Role of Breastfeeding on the Prevention of Infant Mortality. Lancet
52. Gregson S, Terceira N, Kakowa M, et al. Study of bias in antenatal clinic HIV-1 surveillance data in a high contraceptive prevalence population in sub-Saharan Africa. AIDS
53. Central Statistical Office, Zimbabwe. Demographic Health Survey (DHS), 2005-6 Preliminary Report. Harare, Zimbabwe. Available at: http://harare.usembassy.gov/uploads/images/ErYMhCdkkmrjnsEXXPytlA/Zimbabwe_Preliminary-A4_paper.pdf
. Accessed October 5, 2007.
54. Chama C, Gashau W, Oguche S. The value of highly active antiretroviral therapy in the prevention of mother-to-child transmission of HIV. J Obstet Gynaecol
55. Maman S, Mbwambo J, Hogan NM, et al. Women's barriers to HIV-1 testing and disclosure: challenges for HIV-1 voluntary counselling and testing. AIDS Care
56. Sherr L, Lopman B, Kakowa M, et al. Voluntary counselling and testing: uptake, impact on sexual behaviour, and HIV incidence in a rural Zimbabwean cohort. AIDS
57. Bassett MT. Ensuring a public health impact of programs to reduce HIV transmission from mothers to infants: the place of voluntary counseling and testing. Am J Public Health
58. Cunningham CK, Chaix ML, Rekacewicz C, et al. Development of resistance mutations in women receiving standard antiretroviral therapy who received intrapartum nevirapine to prevent perinatal human immunodeficiency virus type 1 transmission: a substudy of Pediatric AIDS Clinical Trials Group Protocol 316. J Infect Dis
59. Eshleman SH, Jackson JB. Nevirapine resistance after single dose prophylaxis. AIDS Rev
60. Jackson JB, Becker-Pergola G, Guay LA, et al. Identification of the K103N resistance mutation in Ugandan women receiving nevirapine to prevent HIV-1 vertical transmission